Skip to content

Search wahloon.net

Search models, news, guides and FAQs.

Model catalog
Llama8BAbliteration~5–6 GB

Meta-Llama-3-8B-Instruct-abliterated-v3

meta-llama/Meta-Llama-3-8B-Instruct

Classic early Llama 3 abliterated model. Still useful on lower-end hardware.

Run it

Ollama

This repository ships GGUF files. Ollama can pull them from Hugging Face; adjust the quant tag if needed:

ollama run hf.co/failspy/Meta-Llama-3-8B-Instruct-abliterated-v3:Q4_K_M

Download the weights

Uses the Hugging Face CLI. Install it with pip install -U huggingface_hub if needed.

hf download failspy/Meta-Llama-3-8B-Instruct-abliterated-v3

LM Studio / llama.cpp

In LM Studio, search for the repository name and pick a GGUF quant that fits your VRAM. With llama.cpp, use llama-cli -hf failspy/Meta-Llama-3-8B-Instruct-abliterated-v3. See the setup guide for full flags and sampling defaults.

Formats & tags

GGUFTransformers

Listed quantizations: GGUF, Transformers.

#llama#8b#classic#abliterated

Notes

Classic early Llama 3 abliterated model. Still useful on lower-end hardware.

Method: abliteration. License: Llama 3. Last catalog update: Dec 1, 2024. This entry links to a third-party repository; wahloon.net is not affiliated with the publisher. Verify the current model card, license and file list before downloading. You are responsible for how you use the model.

Browse all
Llama70B
Abliteration

Llama-3.3-70B-Instruct-abliterated

meta-llama/Llama-3.3-70B-Instruct

Large Llama 3.3 abliterated variant for high-end local setups.

Approx. VRAM
~40+ GB
License
Llama 3.3
Downloads
95K / 30d
#llama#70b#abliterated

updated Sep 20, 2026